Abstract More than one million Americans harbor cerebral cavernous malformations (CCMs), consisting of clusters of proliferating dilated hemorrhagic capillaries, predisposing patients to a lifetime risk of stroke, seizures, and other neurologic sequelae. CCMs can develop sporadically, but the risk to develop CCMs can also be inherited by mutation in one of three recently identified genes. The presence of solitary lesions in sporadic cases and multiple, randomly-scattered lesions in autosomal dominant cases have led us to hypothesize that CCM lesion genesis follows a Knudson two-hit mechanism requiring somatic loss of the wild-type allele, leading to endothelial cell proliferation and CCM growth. We can test our hypothesis by analyzing human CCM lesion tissue for the presence of bi-allelic mutations in the respective CCM gene. Unfortunately, the assessment of somatic mutations and proliferative activity at earliest stages of lesion development in humans is hindered by a strong selection bias in tissue availability. Smaller CCMs, especially at earliest phases of lesion genesis, are rarely if ever surgically resected. Thus, we cannot determine whether bi-allelic mutations identified in the human lesions are the cause or merely an effect of lesion growth. In order to test our two-hit hypothesis for lesion genesis, we have created mouse models with targeted mutations of either Ccm1 or Ccm2, where heterozygote animals develop CCM lesions only in a genetic background null for the tumor suppressor Trp53 (p53) gene. We are concurrently developing second generation models, where the Ccm genes are similarly crossed into a genetic background deficient in mismatch repair, in order to bias somatic mutations towards simple point mutations easily identified by DNA sequencing. We have in parallel developed high resolution magnetic resonance (MR) imaging and stereotactic localization techniques which enable the detection and characterization of murine lesions, and cellular dissection from lesions at earliest stages of development. Using this novel platform and a collection of banked excised human lesions, we propose three specific aims investigating the following hypotheses: (1) that somatic mutation in the wild type copy of the appropriate Ccm gene occurs at early (dilated capillary) stage of lesion genesis, and is expanded clonally in cavernous vessels of later stage (multi-cavernous) CCM lesions in mice with targeted mutagenesis; (2) that lesions develop and grow over time, and endothelial cells in both early and late stage murine CCM lesions exhibit proliferative activity and alterations of the blood-brain-barrier; and (3) that somatic mutations in resected human CCM lesions validate the two-hit mechanisms of lesion genesis in man. This work will shed light on the molecular mechanism of CCM genesis, and aid in our long term goal of finding new approaches to their prevention and treatment.